Much of the craniofacial skeleton arises from the pharyngeal arches, 3D structures that undergo complex changes in shape and gene expression over time. However, detailed analyses of early gene expression profiles in the arches have not been performed in vivo. Without such information, it is impossible to predict how gene expression changes will affect subsequent skeletal development. Our labs have extensively studied the endothelin1 (Edn1), bone morphogenetic protein (Bmp), Wnt and Jag/Notch signaling pathways in the arches, as these four pathways pattern the dorsal-ventral (D-V) axis of the facial skeleton. We have shown that Edn1 and Bmp signaling initially promote ventrally- expressed genes and later subdivide the arches into separate D-V sub-domains. Our preliminary data suggest that Wnt signaling controls competence to respond to Edn1/Bmp and that these three pathways are opposed by Jag/Notch signaling. The pathways regulated by these four signals are highly dynamic, containing multiple feedback loops and crosstalk that create a robust system resistant to perturbation. With a collection of mouse and zebrafish mutants in all four signals, we are in a unique position to assess conservation of gene expression across species in sufficient detail for computational modeling. Our goal is to use these models to predict in silicon facial defects observed following genetic perturbations of these signals. Our dual-species approach will identify new candidate genes and generate models that are clinically relevant to human craniofacial genetics. To address these goals we will pursue two specific aims. In Aim 1, we hypothesize that while Edn1, Bmp, Wnt and Jag/Notch signaling are all critical for establishing the initial identities of skeletal progenitor in the arches along the D-V axis, they each play distinct roles. We will address this hypothesis by using high-throughput RNA sequencing to define early changes in gene expression in mutants of all four pathways. These Early Response Profiles (ERPs) will be used to produce models that integrate gene expression changes across mutants to understand both the unique roles of each factor and crosstalk between signals. In Aim 2 we hypothesize that core sets of enhancers are responsible for the ERP for each signal, some of which mediate crosstalk between or feedback within a signaling pathway, as well as insulating pathways from one another. We will address this by isolating enhancers for genes identified in Aim 1 and testing their activities in both mice and zebrafish. These will be incorporated into our mathematical models to understand enhancer sensitivity and how this regulates sharpness of gene expression boundaries. Our long-term goal is to build a comprehensive model for a craniofacial gene regulatory network that can be amended as new data are available.